Abstract:
|
Some diagnostic tests identify subjects at risk for developing disease. High risk subjects may be treated to prevent or delay disease onset. For example, women screened for cervical dysplasia with cytology and HPV testing are referred to colposcopy, a highly effective procedure for decreasing risk of uterine cervical cancer. If disease incidence is rare or progression to symptomatic disease is slow, certain retrospective studies (case-cohort, nested case-control) are efficient for estimating the predictive value of the test for future disease, assuming analytical performance of the test on stored specimens is acceptable. Clinical utility of a test depends not just on its predictive value but also on the safety and efficacy of available treatment options. Relative utility may be evaluated at the risk threshold, the risk of disease at which a patient is indifferent to treatment or no treatment. The risk threshold is a patient preference. Incorporating patient preferences into medical device evaluation is an FDA/CDRH strategic priority. We'll discuss the described approaches to support regulatory approval of a diagnostic test intended to guide prophylactic treatment decisions.
|